Skip to content

Latest commit

 

History

History
93 lines (59 loc) · 2.09 KB

README.md

File metadata and controls

93 lines (59 loc) · 2.09 KB

Bunny-LLM

a Deno LLM API Service

Usage

Set base_url to https://bunny-llm.deno.dev/v1/ or your own endpoint.

Set env variable BUNNY_API_TOKEN and specific vendor tokens.

Model format [vendor]:[model_name], for example openai:gpt-4-turbo

Example

from openai import OpenAI

client = OpenAI(
    base_url='https://bunny-llm.deno.dev/v1/',
    api_key='YOUR_BUNNY_API_TOKEN',
)
model_name = 'cloudflare:@cf/qwen/qwen1.5-0.5b-chat'

res = client.chat.completions.create(
    model=model_name,
    messages=[
        {'role': 'user', 'content': 'Who are you?'}
    ],
)
print(res.choices[0].message.content)

Free ChatGPT 3.5

vendor is free

api_key is EMPTY

model is gpt-3.5-turbo only

CloudFlare Workers AI

vendor is cf or cloudflare

model can refer to Supported Models

Set env variable CF_ACCOUNT_ID and CF_API_TOKEN or use the following algorithm.

// JavaScript
api_key = encodeURIComponent(JSON.stringify({
    'account': 'YOUR_CF_ACCOUNT_ID',
    'token': 'YOUR_CF_API_TOKEN',
}))
# python
import json
import urllib.parse

api_key=urllib.parse.quote(json.dumps({
    'account': 'YOUR_CF_ACCOUNT_ID',
    'token': 'YOUR_CF_API_TOKEN',
}))

DashScope

vendor is ds or dash_scope

model can refer to Supported Models

api_key can set by DASHSCOPE_API_KEY or pass it directly Get API KEY

Groq

vendor is groq

api_key can set by GROQ_API_KEY or pass it directly Get API KEY

model can refer to Supported Models

Deploy

Vercel

Deploy with Vercel

Netlify

Deploy to Netlify